For example, none of the
product speci?¬?cations used a generalization relationship, a concept taken from MOF.
The modeling tool allowed removing the unused concepts or just hiding them from the
modelers. Particularly useful was that this was possible even when the language was
already used to specify dozens of insurance products.
6.5 GENERATOR FOR JAVA
The target platform included a quotation engine that used the information from the
models in the form of Java code. Later, the Java code was expected to be replaced by
XML. With a DSM solution, this change was not seen as problematic since it would
only require changing the generator. Just one person, the generator developer, would
have to work to make a new generator allowing generation of the desired XML from
the same speci?¬?cation models. The users of the modeling language would not even
notice the change in the generated output!
132 INSURANCE PRODUCTS
The main starting point for the generator de?¬?nition was the expected output: the
code that was written manually before considering a DSM solution. Although the
manually written code was taken as a requirement and used as a test case for the
generator, its structure was not implemented in the generator: the generator produced
the same functionality but in a different structure from the manually written code.
Pages:
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262